I am just starting on with mahout and is thinking of building a recommendation engine from my existing data. The data in my current tables is stored in the format like
userId From TO Meal Carrier Travel type
MT001 London Abu Dhabi Non-Veg Lufthansa International
MT001 Abu Dhabi Beijing Veg Lufthansa International
MT001 New York Chicago Non-Veg American Airlines Domestic
MT002 New York Texas Veg American Airlines Domestic
Now I want to convert this data set to the data model which is accepted by mahout. I know that format for mahout data model is like
userId(int),itemId(int),preference value(float)
Can anyone please suggest how can I convert my travel data set above( which includes much more column and different datatypes than mohout data model) to the data model supported by mahout. Any help will be appreciated.
I guess you have to rewrite you Data model class, try extending and rewriting the AbstractDataModel class,you can also try using IDMigrator control which is provided in the framework to convert your string to integer value.
Go through this topic- Mahout : To read a custom input file